Pumped Storage Power Station Dam Safety Monitoring Method Based on Beidou Positioning

Abstract

The present invention relates to the technical field of precise measurement of geometric quantities based on satellite positioning, and in particular to a method for safety monitoring of pumped-storage power station dams based on Beidou positioning. By deploying Beidou monitoring stations and base station arrays in the dam body and slope areas, multi-band satellite signals are synchronously received and edge computing is performed. Pre-processed observation data is generated by filtering and fusing effective frequency band data using a signal-to-noise ratio threshold, and uploaded to a cloud server in real time via a 4G network and a Beidou short message dual-channel transmission link. The cloud server constructs a local enhancement network, adopts a multi-base station joint error modeling method to correct ionospheric and tropospheric errors, and dynamically optimizes the monitoring point distribution density, signal processing parameters, and atmospheric refraction compensation coefficients through a closed-loop mechanism. Dual-channel redundancy check and interpolation completion technology are combined to ensure data integrity. The present invention realizes dynamic monitoring of millimeter-level deformation of the entire dam under complex terrain, improving the uniformity of temporal and spatial benchmarks, data real-time performance, and structural safety assessment accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of precise measurement of geometric quantities based on satellite positioning, specifically to the technical field of dynamic monitoring of three-dimensional deformation of pumped-storage power station dams using Beidou positioning technology, and in particular to a pumped-storage power station dam safety monitoring method based on Beidou positioning. Background Art

[0002] Existing pumped-storage dam deformation monitoring utilizes precise geometric measurement techniques to obtain real-time displacement response characteristics of the dam structure under the coupled effects of static loads, hydraulic penetration, and temperature fluctuations. Core methods include geodetic surveying, satellite positioning, static leveling, and sensor-embedded monitoring. Geodetic surveying establishes a datum network using total stations or precision levels, periodically observing horizontal displacement and vertical settlement of the dam surface. Satellite positioning utilizes Global Navigation Satellite System (GNSS) receivers to continuously collect the three-dimensional coordinates of key points on the dam crest and analyze dynamic deformation trends. Static leveling monitors differential settlement of the dam body by measuring changes in the liquid level in connecting pipes. Internal deformation monitoring relies on multi-point extensometers, inclinometers, and fiber Bragg grating sensors to measure strain accumulation and crack propagation at different elevations within the dam body. These data, integrated with statistical models and machine learning algorithms, can quantitatively assess dam structural stiffness degradation and stability thresholds, providing a basis for operational risk warnings.

[0003] Existing pumped-storage power station dam safety monitoring faces the following technical pain points in deformation monitoring: Existing manual monitoring methods rely on total stations for surface deformation measurement. Limited by terrain complexity and signal interference, these methods result in insufficient observation frequency, poor real-time performance, and difficulty establishing a unified spatiotemporal benchmark. These methods are unable to meet the requirements for dynamic perception of millimeter-level three-dimensional deformation across the entire dam area. For example, surface deformation monitoring of the dam body and slopes of the upper and lower reservoirs of pumped-storage power stations relies on periodic manual total station measurements due to rugged terrain and electromagnetic interference. These data collection intervals for horizontal displacement and vertical settlement are long, making it impossible to capture the dynamic characteristics of deformation induced by instantaneous load changes or temperature stress. Furthermore, manual data processing carries the risk of cumulative errors due to inconsistent spatiotemporal benchmarks. Summary of the Invention

[0004] In response to the shortcomings of existing technologies, the present invention provides a pumped-storage power station dam safety monitoring method based on Beidou positioning. The present invention solves the problems of insufficient manual observation frequency, inconsistent time and space benchmarks, and inability to obtain millimeter-level three-dimensional deformation dynamic data in real time in the existing pumped-storage power station dam deformation monitoring due to terrain complexity and signal interference.

[0005] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows:

[0006] The present invention provides a method for monitoring the safety of a pumped storage power station dam based on Beidou positioning, comprising:

[0007] Step 1: Deploy Beidou monitoring stations and reference stations in the dam body and slope areas to receive multi-band satellite signals, including Beidou satellite signals, GPS satellite signals, GLONASS satellite signals, and Galileo system satellite signals. Perform edge computing on the received multi-band satellite signals, filter and fuse effective frequency band data through a signal-to-noise ratio threshold, generate pre-processed satellite observation data, and establish a dual-channel transmission link through a preset 4G network and a preset Beidou short message to upload the data to a preset cloud server in real time.

[0008] Step 2: Build a local augmentation network on the cloud server, perform dynamic error correction on the uploaded pre-processed satellite observation data, generate three-dimensional coordinate solution results, and output the calibrated spatiotemporal reference data stream;

[0009] Step 3: performing spatiotemporal alignment on the satellite displacement data in the calibrated spatiotemporal reference data stream with the strain measurement data collected by the static level and the strain measurement data collected by the multi-point displacement meter. By fusion of the data, the transient response characteristics of the displacement and strain are extracted, and a global deformation dynamic model that integrates the coupled effects of the temperature field and the seepage field is constructed.

[0010] Step 4: Input the satellite displacement data of the spatiotemporal reference data stream and the strain measurement data into the global deformation dynamic model, and output the displacement, velocity, and acceleration monitoring parameters of the dam structure;

[0011] Step 5: Setting warning thresholds corresponding to the displacement, velocity, and acceleration monitoring parameters of the dam structure on the cloud server. When the displacement, velocity, and acceleration monitoring parameters of the dam structure exceed the corresponding warning thresholds, generating a graded warning signal including a risk level and location coordinates based on the values ​​of the exceeded parameters;

[0012] Step 6: construct a three-dimensional deformation thermodynamic map based on the displacement, velocity and acceleration monitoring parameters of the dam structure, and push a risk report including the coordinate positioning of the thermodynamic map to a pre-configured terminal device through a preset communication link.

[0013] Furthermore, in the Beidou positioning-based pumped storage power station dam safety monitoring method of the present invention, step 2 includes:

[0014] Beidou monitoring stations and reference stations deployed in the dam and slope areas synchronously receive the Beidou, GPS, GLONASS and Galileo satellite signals;

[0015] Perform edge computing processing on the received multi-system satellite signals, filter through signal-to-noise ratio thresholds to retain valid frequency band data, and add signal quality identifiers;

[0016] Predicting a communication blind spot based on a signal quality identifier, and switching to a Beidou short message transmission mode in the dual-channel transmission link when the signal strength is detected to be lower than a preset strength threshold;

[0017] Upload the pre-processed satellite observation data to the cloud server through the preset 4G network and Beidou short message dual channel;

[0018] The pre-processed data is temporally and spatially aligned on the cloud server to generate a continuous data stream for input into step 2 to perform dynamic error correction.

[0019] Furthermore, in the Beidou positioning-based pumped storage power station dam safety monitoring method of the present invention, step 2 includes:

[0020] A BeiDou local augmentation network is constructed by deploying base stations, and a multi-base station joint error modeling method is used to correct ionospheric and tropospheric errors in real time for pre-processed satellite observation data.

[0021] Sub-nanosecond time synchronization is achieved based on the high-precision time synchronization device preset at the base station, and positioning deviation is eliminated through the combined processing of baseline solution and network adjustment;

[0022] Input the corrected satellite observation data into the three-dimensional coordinate solution processing flow to generate a spatiotemporal reference data stream with a timestamp;

[0023] The space-time reference data stream is recalibrated using preset atmospheric refractometer compensation parameters, and the calibrated space-time reference data stream is output.

[0024] Furthermore, in the pumped storage power station dam safety monitoring method based on Beidou positioning described in the present invention, step 3 includes: extracting satellite displacement data from the calibrated time-space reference data stream, and aligning the time stamps with the strain data collected by the static level and the strain data collected by the multi-point displacement meter;

[0025] The preset Kalman filter algorithm is used to perform cross-dimensional fusion of the aligned displacement and strain data to extract horizontal displacement, vertical settlement and crack propagation characteristics;

[0026] The fused data is combined with the field data collected by the temperature sensor and the seepage pressure gauge and input into the global deformation dynamic model;

[0027] The transient response consistency between the fused data and the global deformation dynamic model output is verified through time series analysis, and the transient response consistency verification results are generated for early warning rules to call.

[0028] Furthermore, in the pumped storage power station dam safety monitoring method based on Beidou positioning described in the present invention, step 4 includes: establishing a stiffness degradation model on a cloud server based on the displacement parameters output by the global deformation dynamic model and setting multi-dimensional warning thresholds for deformation, velocity, and acceleration;

[0029] Compare the displacement monitoring value of the deformation dynamic model with the warning threshold in real time, and trigger a graded warning signal when the monitoring value exceeds the limit for a set number of consecutive times;

[0030] The coordinates of the out-of-limit area are mapped to the 3D deformation thermal map through the preset digital twin model, and the transient response consistency verification results are superimposed to obtain the marked high-risk areas.

[0031] The marked high-risk areas are pushed to the preset inspection terminal, and the warning threshold parameters are updated based on the disposal results fed back by the inspection terminal.

[0032] Furthermore, in the pumped storage power station dam safety monitoring method based on Beidou positioning according to the present invention, step 5 includes:

[0033] Collect real-time displacement monitoring information based on the deployed base stations, and eliminate system errors caused by base station structural deformation through a preset dynamic coordinate update algorithm;

[0034] Based on the historical monitoring feature distribution law generated by the spatiotemporal alignment and data fusion processing, the spatial distribution density of the initially deployed monitoring points and the data acquisition frequency are dynamically optimized, wherein the historical monitoring feature distribution law is obtained by the time series analysis of the transient response consistency verification results;

[0035] Dynamically optimize the noise suppression coefficient and multi-band signal weight matching parameters in the signal-to-noise ratio threshold screening process based on the data quality evaluation index generated by the secondary calibration process of the atmospheric refractometer compensation parameter;

[0036] Combined with the environmental parameter change data in the graded warning signal, the atmospheric refractometer compensation parameters described in step 3 are updated according to the preset calibration cycle.

[0037] Furthermore, in the pumped storage power station dam safety monitoring method based on Beidou positioning according to the present invention, step 6 includes:

[0038] When the 4G network signal strength is higher than the set signal strength threshold, the pre-processed satellite observation data generated in step 1 is uploaded through the high-speed channel;

[0039] When it is detected that the network delay exceeds the preset tolerance threshold or the data packet loss rate increases, switching to the Beidou short message transmission mode in the dual-channel transmission link to send compressed data packets;

[0040] The cloud server performs redundancy check and timestamp alignment on the dual-channel transmission link data based on the signal quality identifier added in step 2;

[0041] The reorganized data is transmitted to the spatiotemporal reference data stream update processing flow described in step 3, and a linear interpolation completion operation is performed on the low-quality data segments marked by the identifier.

[0042] Beneficial effects of the present invention:

[0043] The present invention improves the reliability of data acquisition under complex terrain through multi-system satellite signal fusion and dynamic error correction technology. The Beidou monitoring station and the base station array are combined with a dual-channel transmission link to ensure the continuity of data transmission in communication blind spots. The construction of a local enhanced network and multi-base station joint error modeling eliminate interference between the ionosphere and the troposphere, generate a time-space reference data stream with millimeter-level precision, and realize the unification of the time-space reference of multi-source heterogeneous data; cross-dimensional data fusion based on the Kalman filter algorithm extracts horizontal displacement, vertical settlement and crack extension characteristics, and combines the temperature field and seepage field coupling modeling to construct a global deformation dynamic model, and realizes dynamic mapping and visual early warning of high-risk area coordinates through digital twin technology; the closed-loop adaptive mechanism dynamically optimizes the monitoring point distribution density and signal processing parameters according to historical monitoring characteristics, and combines the environment-driven atmospheric refraction compensation update and dual-link redundant verification to form a full-process closed-loop management from data acquisition, error correction to graded early warning, effectively improving the real-time and accuracy of dam deformation monitoring and the reliability of structural safety assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on the drawings without paying any creative labor.

[0045] Figure 1 This is a flowchart of a pumped storage power station dam safety monitoring method based on Beidou positioning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. The technical solutions provided by each embodiment of the present invention are described in detail below in conjunction with the drawings. In order to better understand the purpose of the present invention, the present invention is further described in detail below.

[0047] See also Figure 1 The present invention provides a pumped storage power station dam safety monitoring method based on Beidou positioning, comprising:

[0048] Step 1: Deploy Beidou monitoring stations and reference stations in the dam body and slope areas to receive multi-band satellite signals, including Beidou satellite signals, GPS satellite signals, GLONASS satellite signals, and Galileo system satellite signals. Perform edge computing on the received multi-band satellite signals, filter and fuse effective frequency band data through a signal-to-noise ratio threshold, generate pre-processed satellite observation data, and establish a dual-channel transmission link through a preset 4G network and a preset Beidou short message to upload the data to a preset cloud server in real time.

[0049] Step 2: Build a local augmentation network on the cloud server, perform dynamic error correction on the uploaded pre-processed satellite observation data, generate three-dimensional coordinate solution results, and output the calibrated spatiotemporal reference data stream;

[0050] Step 3: performing spatiotemporal alignment on the satellite displacement data in the calibrated spatiotemporal reference data stream with the strain measurement data collected by the static level and the strain measurement data collected by the multi-point displacement meter. By fusion of the data, the transient response characteristics of the displacement and strain are extracted, and a global deformation dynamic model that integrates the coupled effects of the temperature field and the seepage field is constructed.

[0051] Step 4: Input the satellite displacement data of the spatiotemporal reference data stream and the strain measurement data into the global deformation dynamic model, and output the displacement, velocity, and acceleration monitoring parameters of the dam structure;

[0052] Step 5: Setting warning thresholds corresponding to the displacement, velocity, and acceleration monitoring parameters of the dam structure on the cloud server. When the displacement, velocity, and acceleration monitoring parameters of the dam structure exceed the corresponding warning thresholds, generating a graded warning signal including a risk level and location coordinates based on the values ​​of the exceeded parameters;

[0053] Step 6: construct a three-dimensional deformation thermodynamic map based on the displacement, velocity and acceleration monitoring parameters of the dam structure, and push a risk report including the coordinate positioning of the thermodynamic map to a pre-configured terminal device through a preset communication link.

[0054] The specific technical solution for the Beidou-based dam safety monitoring method for pumped-storage power stations is as follows: Beidou monitoring stations and base stations are deployed within the dam and slope areas to receive multi-band satellite signals from Beidou, GPS, GLONASS, and Galileo. These multi-band satellite signals are processed in real time through edge computing. Signal-to-noise ratio threshold filtering is used to fuse data from valid frequency bands, eliminate interference from low-SNR frequency bands, and generate pre-processed satellite observation data with signal quality identifiers. This pre-processed data is uploaded to a cloud server in real time via a pre-set dual-channel transmission link using the 4G network and Beidou short message transmission. The Beidou short message transmission mode is automatically activated when the signal strength falls below a preset threshold, ensuring continuous data transmission in communication blind spots.

[0055] The cloud-based server constructs the Beidou local area augmentation network and employs a multi-base station joint error modeling approach to perform dynamic error correction on uploaded pre-processed satellite observation data. Sub-nanosecond time synchronization is achieved through the high-precision time synchronization devices of the base station cluster. A combined baseline solution and network adjustment process eliminates positioning errors caused by ionospheric delay, tropospheric refraction, and multipath effects. The corrected satellite observation data is input into the three-dimensional coordinate solution process to generate a time-stamped space-time reference data stream. This data is then recalibrated using preset atmospheric refraction compensation parameters, outputting a space-time reference data stream with millimeter-level accuracy.

[0056] Satellite displacement data from the calibrated spatiotemporal reference data stream are time-stamp aligned with strain measurement data collected by a static level and multi-point displacement gauges. A Kalman filter algorithm is then used to perform cross-dimensional fusion of these multi-source data. Characteristic parameters for horizontal displacement, vertical settlement, and crack propagation are extracted during the fusion process. Combined with temperature and seepage field data collected by temperature sensors and seepage pressure gauges, a global deformation dynamic model incorporating multi-physics coupling effects is constructed. This model verifies the transient response consistency between the fused data and the model output through time series analysis, generating transient response consistency verification results.

[0057] The displacement, velocity, and acceleration monitoring parameters output by the global deformation dynamic model are fed into the early warning analysis module on the cloud server, where multi-dimensional warning thresholds corresponding to the structural safety level are set. When a monitoring parameter exceeds a threshold, a graded warning signal is triggered based on the number and value of the exceeded parameters. The coordinates of the exceeded areas are mapped to a three-dimensional deformation heat map using the digital twin model. The heat map is overlaid with the transient response consistency verification results, annotating the coordinates of high-risk areas and sending them to the inspection terminal, forming a closed-loop response process.

[0058] Monitoring point density and observation frequency are dynamically optimized based on the distribution characteristics of historical monitoring data. Noise suppression parameters and signal weighting are dynamically adjusted based on the quality assessment results of the spatiotemporal benchmark data stream. Systematic errors caused by base station deformation are corrected in real time using a dynamic coordinate update algorithm. Atmospheric refraction compensation parameters are regularly updated based on environmental parameter changes in early warning records to maintain consistent monitoring accuracy.

[0059] The dual-channel transmission link adaptively switches transmission modes based on network conditions. When 4G network signal strength exceeds a set threshold, data is transmitted using the high-speed channel. When network latency or packet loss exceeds a certain limit, data is transmitted using Beidou short messages, which transmit compressed data packets. The cloud server performs redundancy checks and timestamp alignment on the dual-channel data based on signal quality identifiers. The reassembled data stream is then updated to the spatiotemporal reference data stream processing flow. Low-quality data segments marked by the identifiers are then linearly interpolated to ensure data integrity and timeliness.

[0060] The above technical solution realizes millimeter-level dynamic monitoring of the entire dam deformation through multi-band signal fusion, dynamic error correction, multi-source data fusion and closed-loop early warning mechanism, solves the technical problems of insufficient frequency of manual observation and inconsistent time and space benchmarks in complex terrain, and improves the real-time and accuracy of structural safety early warning.

[0061] Specifically, in the pumped storage power station dam safety monitoring method based on Beidou positioning according to the present invention, step 2 includes:

[0062] Beidou monitoring stations and reference stations deployed in the dam and slope areas synchronously receive the Beidou, GPS, GLONASS and Galileo satellite signals;

[0063] Perform edge computing processing on the received multi-system satellite signals, filter through signal-to-noise ratio thresholds to retain valid frequency band data, and add signal quality identifiers;

[0064] Predicting a communication blind spot based on a signal quality identifier, and switching to a Beidou short message transmission mode in the dual-channel transmission link when the signal strength is detected to be lower than a preset strength threshold;

[0065] Upload the pre-processed satellite observation data to the cloud server through the preset 4G network and Beidou short message dual channel;

[0066] The pre-processed data is temporally and spatially aligned on the cloud server to generate a continuous data stream for input into step 2 to perform dynamic error correction.

[0067] The specific technical implementation process for Step 2 is as follows: An array of Beidou monitoring stations and reference stations is deployed at a predetermined spatial density within the dam and slope areas. Each station simultaneously receives multi-band satellite signals from Beidou, GPS, GLONASS, and Galileo. The monitoring stations are equipped with built-in multi-mode GNSS receivers, which capture navigation signals from different satellite systems in parallel through multi-frequency RF channels. Carrier phase difference technology is used to enhance the accuracy of raw observation data. The reference stations are equipped with high-stability pedestals and anti-interference shielding enclosures. Submillisecond clock synchronization with the monitoring stations is achieved via optical fiber transmission, ensuring consistent temporal and spatial references for multi-source satellite signal acquisition.

[0068] Received multi-system satellite signals are transmitted to edge computing nodes for real-time processing. The edge computing nodes integrate a signal quality assessment module and a frequency band fusion algorithm. The signal quality assessment module dynamically calculates the carrier-to-noise ratio (CNR) parameters for each frequency band based on a signal-to-noise ratio (SNR) threshold screening mechanism, eliminating low-quality frequency band data below a set threshold. Valid frequency band data is then integrated using a weighted fusion algorithm. Fusion weights are dynamically assigned based on the geometric dilution of precision (DOP) of each satellite system's constellation distribution. This generates pre-processed satellite observation data with a signal quality identifier. The signal quality identifier, which includes the CNR, multipath effect index, and data integrity check code, is used for quality control of subsequent transmission links.

[0069] A signal quality monitoring module is built into the cloud server, which analyzes the signal quality identifiers in preprocessed data in real time and predicts communication blind spots based on carrier-to-noise ratio degradation trends and multipath effect indices. If the signal strength falls below a preset threshold or if the data integrity check fails, a transmission mode switch is automatically triggered. Data transmission continuity is maintained during this switchover process, with compressed data packets preferentially sent via the Beidou short message transmission mode. The compression algorithm utilizes entropy coding to preserve the phase and pseudorange information of the original observation data, preventing the loss of critical data.

[0070] Preprocessed satellite observation data is uploaded in parallel via a pre-configured dual-channel transmission link, the 4G network, and the Beidou short message channel. Both channels utilize differentiated quality of service strategies. The 4G network channel carries the high-frequency raw observation data stream, while the Beidou short message channel transmits compressed data packets for redundancy. The cloud server's data receiving module aligns timestamps and reassembles data packets on both channels, eliminating timing misalignment caused by transmission delays.

[0071] In the cloud server's data processing engine, preprocessed data is input into the spatiotemporal alignment module for refined processing. This module uses a sliding window mechanism to match satellite displacement data with the reference station coordinate frame. A least-squares adjustment algorithm is used to eliminate spatial deviations caused by receiver clock and orbit errors. The aligned data stream is interpolated and resampled according to a unified time base, generating a standardized dataset with a continuous time series. This data is then input into the dynamic error correction process for ionospheric and tropospheric error modeling. This process utilizes a local augmentation network constructed from a cluster of reference stations. Through multi-base station joint calculations, common mode errors and random noise are separated in real time, resulting in millimeter-level accuracy in the output of three-dimensional coordinate solutions.

[0072] The above technical solution forms a complete satellite observation data preprocessing and error correction chain through multi-system signal fusion, edge quality screening, dual-channel redundant transmission and unified space-time benchmark processing, providing high-precision space-time benchmark data support for subsequent deformation dynamic modeling.

[0073] Specifically, in the pumped storage power station dam safety monitoring method based on Beidou positioning according to the present invention, step 2 includes:

[0074] A BeiDou local augmentation network is constructed by deploying base stations, and a multi-base station joint error modeling method is used to correct ionospheric and tropospheric errors in real time for pre-processed satellite observation data.

[0075] Sub-nanosecond time synchronization is achieved based on the high-precision time synchronization device preset at the base station, and positioning deviation is eliminated through the combined processing of baseline solution and network adjustment;

[0076] Input the corrected satellite observation data into the three-dimensional coordinate solution processing flow to generate a spatiotemporal reference data stream with a timestamp;

[0077] The space-time reference data stream is recalibrated using preset atmospheric refractometer compensation parameters, and the calibrated space-time reference data stream is output.

[0078] The technical implementation process of step 2 is as follows: a group of base stations are deployed around the dam body according to a preset topological structure to form a Beidou local augmentation network. The base stations are equipped with multi-band GNSS receivers and meteorological sensors to collect satellite raw observation data and atmospheric parameters in real time. The network adopts a distributed data processing architecture, and each base station is interconnected through a fiber-optic dedicated network to achieve millisecond-level synchronous transmission of observation data. The multi-base station joint error modeling method is based on the spatial distribution characteristics of the base station group, constructs a regional differential model of ionospheric delay and tropospheric refraction error, and eliminates systematic errors in the satellite signal propagation path through real-time differential correction technology. A sliding window adaptive filtering algorithm is introduced in the error correction process to dynamically separate the common mode error and site-specific noise in the observation data.

[0079] The base station's built-in rubidium atomic clock and satellite common-view module form a high-precision time synchronization system, achieving sub-nanosecond time synchronization between stations. Time synchronization data is embedded in the observation data stream and serves as the reference input for baseline solution. The baseline solution utilizes a double-difference carrier phase observation model, combined with a network adjustment algorithm to eliminate positioning bias caused by receiver clock errors, orbit errors, and multipath effects. The network adjustment incorporates robust estimation theory, optimizing station coordinate residuals through iterative weighted least squares optimization to output relative positioning results that meet millimeter-level accuracy requirements.

[0080] Corrected satellite observation data is fed into a 3D coordinate calculation module, which integrates a precise point positioning algorithm and a dynamic filtering model. The data processing process employs an extended Kalman filter to fuse pseudorange and carrier phase observations, calculating the 3D geocentric coordinates of the monitoring station in real time. The calculated results are superimposed with the reference station's coordinate frame conversion parameters to generate a displacement data stream in a unified spatiotemporal reference. This data stream is embedded with Coordinated Universal Time (UTC) timestamps, forming a standardized dataset with temporal and spatial correlations for subsequent deformation modeling.

[0081] The spatiotemporal reference data stream is fed into the atmospheric refraction compensation module for secondary calibration. This module utilizes a pre-set atmospheric refraction parameter library, built using historical meteorological data and machine learning algorithms. This library contains signal propagation correction coefficients for various pressure, temperature, and humidity conditions. The calibration process utilizes adaptive weighted interpolation technology, dynamically matching optimal compensation parameters based on real-time meteorological observations to perform nonlinear corrections on the elevation component of the spatiotemporal reference data stream. The calibrated data stream undergoes integrity verification before output, eliminating systematic measurement biases caused by atmospheric environmental variations.

[0082] This approach, through a multi-stage process involving the construction of a local augmentation network, time synchronization optimization, coordinate solution workflow, and atmospheric compensation calibration, forms a complete correction chain from raw observations to a high-precision spatiotemporal benchmark data stream. Collaborative error modeling across a cluster of benchmark stations improves regional differential accuracy, a time synchronization system ensures spatiotemporal consistency of multi-source data, and a dynamic matching mechanism for atmospheric refraction parameters enhances measurement robustness in complex environments, providing fundamental data support for global dynamic deformation modeling.

[0083] Specifically, in the pumped storage power station dam safety monitoring method based on Beidou positioning described in the present invention, step 3 includes: extracting satellite displacement data from the calibrated time-space reference data stream, and aligning the time stamps with the strain data collected by the static level and the strain data collected by the multi-point displacement meter;

[0084] The preset Kalman filter algorithm is used to perform cross-dimensional fusion of the aligned displacement and strain data to extract horizontal displacement, vertical settlement and crack propagation characteristics;

[0085] The fused data is combined with the field data collected by the temperature sensor and the seepage pressure gauge and input into the global deformation dynamic model;

[0086] The transient response consistency between the fused data and the global deformation dynamic model output is verified through time series analysis, and the transient response consistency verification results are generated for early warning rules to call.

[0087] The specific technical implementation process for step 3 is as follows: Satellite displacement data is parsed from the calibrated spatiotemporal reference data stream output from step 2, and a timestamp parser is used to extract the satellite positioning coordinate sequence. The static level and multi-point displacement gauge upload strain measurement data via an industrial bus protocol. The data acquisition module has a built-in high-precision clock source that generates nanosecond timestamps synchronized with the Beidou timing signal. The spatiotemporal alignment module uses a sliding window matching algorithm to align the time bases of the multi-source data. Cubic spline interpolation is performed on the unevenly sampled data to eliminate timing deviations caused by differences in sensor acquisition frequencies.

[0088] The aligned displacement and strain data are fed into a cross-dimensional data fusion engine. A pre-configured Kalman filter algorithm constructs a state-space model, using satellite displacement data as the system observation variable and strain data as a noise compensation term. During the iterative filtering process, multi-dimensional observations, including horizontal displacement, vertical displacement, and crack aperture, are fused. The covariance matrix dynamically adjusts the fusion weights for each dimension, extracting characteristic parameters representing the horizontal displacement component, vertical settlement gradient, and crack growth rate that characterize the dam's deformation.

[0089] The fused multi-source data is input into the global deformation dynamic model pre-processing module, where it is combined with the temperature gradient and seepage pressure field data collected by the temperature sensor array and seepage pressure gauge network deployed within the dam. The temperature field data is mapped into an equivalent deformation correction using the thermodynamic expansion coefficient conversion module. The seepage field data is then used to calculate the weight of the influence of seepage force on structural strain using Darcy's law, forming a multi-physics field coupling input parameter set.

[0090] The global deformation dynamic model is constructed based on a hybrid architecture of finite element theory and data-driven design, employing an offline training and online update mechanism. The model input integrates and fuses multi-source data, and the output generates displacement and stress fields and crack evolution predictions for the dam structure. The time series analysis module performs sliding average and residual analysis on the transient response data output by the model, calculating the correlation coefficient and root mean square error between the measured data and the predicted results, and generating transient response consistency verification indicators. The verification results are transmitted to the early warning rule library via a standardized interface, triggering threshold comparison and risk assessment logic to support the generation and optimization of graded early warning signals.

[0091] This approach establishes a complete analytical chain from raw data to deformation assessment through a progressive process involving spatiotemporal benchmark alignment, multi-dimensional data fusion, multi-physics coupling modeling, and response consistency verification. The Kalman filter algorithm improves the fusion accuracy of multi-source heterogeneous data, the multi-physics coupling mechanism enhances the model's environmental adaptability, and time series analysis ensures the reliability of the model output, providing a high-confidence decision-making basis for structural safety warnings.

[0092] Specifically, in the pumped storage power station dam safety monitoring method based on Beidou positioning described in the present invention, step 4 includes: establishing a stiffness degradation model on a cloud server based on the displacement parameters output by the global deformation dynamic model and setting multi-dimensional warning thresholds for deformation, velocity, and acceleration;

[0093] Compare the displacement monitoring value of the deformation dynamic model with the warning threshold in real time, and trigger a graded warning signal when the monitoring value exceeds the limit for a set number of consecutive times;

[0094] The coordinates of the out-of-limit area are mapped to the 3D deformation thermal map through the preset digital twin model, and the transient response consistency verification results are superimposed to obtain the marked high-risk areas.

[0095] The marked high-risk areas are pushed to the preset inspection terminal, and the warning threshold parameters are updated based on the disposal results fed back by the inspection terminal.

[0096] The specific technical implementation process for Step 4 is as follows: In the structural safety assessment module on the cloud server, a stiffness degradation model is constructed based on the displacement parameters output by the global deformation dynamic model. This stiffness degradation model integrates the constitutive relationship database of dam structural materials. A machine learning algorithm is used to train the mapping relationship between structural stiffness and displacement parameters. Combined with historical load spectrum data, multi-dimensional warning thresholds for deformation, deformation rate, and acceleration are established. The thresholds are set using a dynamic adjustment mechanism, generating graded safety intervals based on dam design parameters, real-time environmental loads, and material aging coefficients, forming a warning threshold system linked to the structural health status.

[0097] The real-time comparison of the displacement monitoring values ​​of the deformation dynamic model against the warning thresholds is achieved using a sliding window statistical method. The monitoring data is fed into an adaptive filtering module for noise suppression, extracting effective deformation features before entering the threshold comparison engine. The comparison engine utilizes an event-driven architecture. When the monitoring value exceeds the corresponding dimension threshold a certain number of times within a continuous sampling period, a hierarchical warning signal is generated. The warning signal includes the dimension identifier, magnitude, and duration of the violation, and is transmitted to the risk analysis module for prioritization.

[0098] The digital twin model is constructed based on the dam's 3D geometric model and finite element mesh. Using a physical field coupling engine, the coordinates of the over-limit areas are mapped to a 3D deformation heat map. The heat map rendering module uses color gradient encoding technology to dynamically adjust the regional color temperature based on the magnitude and rate of displacement overruns. The transient response consistency verification results generated in step 3 are simultaneously overlaid. The overlaid data is visualized using a transparency fusion algorithm to visualize the risk hierarchy. The coordinates of the annotated high-risk areas are matched to the engineering drawing coordinate system using a spatial indexing algorithm to generate a risk location report containing geographic information encoding.

[0099] The marked high-risk area data is pushed to the inspection terminal via a low-latency communication protocol. The terminal integrates an augmented reality navigation module to plan the inspection route based on the coordinate information in the risk location report. The treatment results are transmitted back to the cloud server via the terminal feedback interface. The feedback data includes the deformation characteristics verified on-site, the treatment measures, and the corrected measurement values. The warning threshold parameter update module uses an incremental learning algorithm to dynamically optimize the weight coefficients of the threshold generation model based on the feedback data, improving the accuracy of matching the threshold settings with the actual damage state of the structure.

[0100] This solution forms a closed-loop management chain from deformation monitoring to risk management through stiffness degradation modeling, dynamic threshold comparison, digital twin visualization, and feedback optimization mechanisms. The stiffness degradation model establishes a quantitative correlation between structural performance and deformation parameters, digital twin technology enables spatial location and visualization of risks, and the feedback mechanism ensures adaptive optimization of the early warning system, providing technical support for the full lifecycle safety management of pumped-storage dams.

[0101] Specifically, in the pumped storage power station dam safety monitoring method based on Beidou positioning according to the present invention, step 5 includes:

[0102] Collect real-time displacement monitoring information based on the deployed base stations, and eliminate system errors caused by base station structural deformation through a preset dynamic coordinate update algorithm;

[0103] Based on the historical monitoring feature distribution law generated by the spatiotemporal alignment and data fusion processing, the spatial distribution density of the initially deployed monitoring points and the data acquisition frequency are dynamically optimized, wherein the historical monitoring feature distribution law is obtained by the time series analysis of the transient response consistency verification results;

[0104] Dynamically optimize the noise suppression coefficient and multi-band signal weight matching parameters in the signal-to-noise ratio threshold screening process based on the data quality evaluation index generated by the secondary calibration process of the atmospheric refractometer compensation parameter;

[0105] Combined with the environmental parameter change data in the graded warning signal, the atmospheric refractometer compensation parameters described in step 3 are updated according to the preset calibration cycle.

[0106] The specific technical implementation process for Step 5 is as follows: The base station's built-in multi-mode GNSS receiver collects displacement monitoring data in real time. A pre-set dynamic coordinate update algorithm eliminates systematic errors caused by base station deformation. This dynamic coordinate update algorithm employs a robust Kalman filter model, integrating data from the base station's structural health monitoring sensors with satellite positioning observations to construct a three-dimensional deformation compensation matrix for the base station. This compensation matrix is ​​dynamically updated using a sliding window mechanism, correcting for deviations in the base station's coordinate frame in real time and improving the baseline stability of the local area augmentation network.

[0107] The historical monitoring data generated through spatiotemporal alignment and data fusion is fed into the feature analysis engine. Using a time series clustering algorithm, the engine extracts the distribution patterns of deformation hotspots. Spatial correlation analysis is performed on this distribution pattern and the transient response consistency verification results from step 4 to identify highly sensitive monitoring areas. The monitoring point optimization module dynamically adjusts monitoring point density based on this analysis, adding redundant sensor nodes in highly sensitive areas and reducing data collection frequency in areas of low change, achieving adaptive allocation of monitoring resources.

[0108] The data quality assessment metrics generated by the secondary calibration of atmospheric refraction compensation parameters are fed into the parameter optimization module. These metrics include the signal-to-noise ratio attenuation gradient, the multipath effect index, and the standard deviation of the phase residual. This optimization module uses a gradient descent algorithm to dynamically adjust the noise suppression coefficient during the signal-to-noise ratio threshold screening process. It also updates the multi-band signal weighting parameters in real time based on changes in the satellite constellation geometry, improving the signal fusion accuracy of edge computing nodes.

[0109] Environmental parameter changes in graded warning signals trigger an update mechanism for atmospheric refractometer compensation parameters. Real-time environmental data collected by temperature, pressure, and humidity sensors is fed into a parameter prediction model. This model, based on a long-short-term memory (LSTM) network, predicts trends in atmospheric refractometer effects. The predictions are combined with a pre-set calibration cycle. When the environmental parameter change exceeds a set threshold or reaches the calibration cycle, the atmospheric refractometer parameter library from step 3 is automatically invoked for an online update. Compensation parameters are then synchronized to all reference stations via a differential broadcast protocol.

[0110] This solution builds a closed-loop adaptive monitoring system through dynamic correction of reference station errors, adaptive optimization of monitoring points, real-time adjustment of signal processing parameters, and environmentally driven updates of atmospheric parameters. Improved reference station stability ensures the accuracy benchmark of the local area augmentation network, while monitoring point optimization enables efficient resource allocation. Dynamic adjustment of signal processing parameters enhances data quality in complex environments, and an environmentally driven compensation mechanism maintains long-term monitoring consistency. Together, these factors contribute to the high accuracy and robustness of the dam safety monitoring system.

[0111] Specifically, in the pumped storage power station dam safety monitoring method based on Beidou positioning according to the present invention, step 6 includes:

[0112] When the 4G network signal strength is higher than the set signal strength threshold, the pre-processed satellite observation data generated in step 1 is uploaded through the high-speed channel;

[0113] When it is detected that the network delay exceeds the preset tolerance threshold or the data packet loss rate increases, switching to the Beidou short message transmission mode in the dual-channel transmission link to send compressed data packets;

[0114] The cloud server performs redundancy check and timestamp alignment on the dual-channel transmission link data based on the signal quality identifier added in step 2;

[0115] The reorganized data is transmitted to the spatiotemporal reference data stream update processing flow described in step 3, and a linear interpolation completion operation is performed on the low-quality data segments marked by the identifier.

[0116] The technical implementation process for Step 5 is as follows: The base station is equipped with a multi-source sensor array to collect structural deformation data in real time. The robust filtering model in the dynamic coordinate update algorithm eliminates its own deformation errors. This algorithm integrates the measurements of the base station's internal strain gauges and inclinometers with the differences in satellite positioning coordinates to construct a three-dimensional deformation compensation matrix for the base station. A sliding window mechanism is used to update the base station's coordinate frame parameters to maintain the stability of the measurement baseline within the local augmentation network.

[0117] After spatiotemporal alignment and data fusion, historical monitoring data is input into the Feature Distribution Analysis module to extract deformation hotspots. This module uses a time series clustering algorithm to identify periods of high variance and spatially clustered areas. Combined with the transient response consistency verification results generated in step 4, this module performs spatial correlation modeling and outputs a monitoring point optimization strategy. This optimization strategy drives the Monitoring Network Reconfiguration module to dynamically adjust the sensor node density, adding redundant monitoring points in high-risk areas and reducing data collection frequency in areas with low-frequency changes, thereby achieving adaptive allocation of monitoring resources.

[0118] The data quality assessment metrics generated by the secondary calibration of atmospheric refraction compensation parameters include carrier phase residual, multipath effect index, and signal-to-noise ratio degradation. These metrics are fed into the parameter optimization engine, which uses a gradient descent algorithm to dynamically adjust the noise suppression coefficient of edge computing nodes. The engine also optimizes the weighting of multi-band signals based on the real-time satellite constellation geometry, improving the anti-interference capability and data validity of the signal preprocessing stage.

[0119] The environmental parameter monitoring module collects temperature, pressure, and humidity data in real time and feeds it into the atmospheric refraction prediction model. The model uses a long-short-term memory network to analyze the nonlinear relationship between environmental parameters and atmospheric refraction. The prediction results trigger a parameter update mechanism. When the magnitude of environmental change exceeds a set threshold or reaches the preset calibration period, the atmospheric refraction parameter library from step 3 is invoked for an online update. The updated parameters are broadcast to all reference stations via a differential correction protocol, maintaining consistent measurement accuracy under complex meteorological conditions.

[0120] This solution forms a closed-loop, adaptive monitoring accuracy control system through reference station error correction, dynamic monitoring network optimization, adaptive signal processing parameter adjustment, and environmentally driven compensation updates. Reference station stability maintenance ensures spatial benchmark uniformity, monitoring point optimization improves data collection efficiency, dynamic signal processing parameter adjustment enhances anti-interference capabilities, and an environmentally coupled parameter update mechanism ensures long-term monitoring stability. Together, these mechanisms support the high accuracy and reliability of the dam safety monitoring system.

[0121] The technical features of the present invention are explained as follows:

[0122] Beidou monitoring station and base station array: a group of Beidou positioning devices deployed in the dam body and slope areas.

[0123] Monitoring stations collect BeiDou satellite signals in real time, while base stations serve as reference points for high-precision positioning. Together, they enhance regional positioning consistency. Equipped with high-performance atomic clocks and anti-interference devices, base stations achieve sub-nanosecond time synchronization via a fiber-optic synchronization network, eliminating clock bias between multiple sites.

[0124] Multi-band satellite signal fusion (Beidou / GPS / GLONASS / Galileo): Simultaneously receives satellite signals from multiple navigation systems, enhancing positioning data reliability through frequency fusion. Fusion of multi-system signals enhances satellite geometry and reduces the risk of signal obstruction or interference from individual systems, ensuring data availability, especially in complex terrain. Edge computing nodes dynamically remove low-quality frequency bands based on the signal-to-noise ratio, retaining valid bands and generating pre-processed data through weighted fusion.

[0125] Edge computing and signal-to-noise ratio threshold screening: Initial signal processing is performed at the data collection end (monitoring station), rather than relying on the cloud. The carrier-to-noise ratio (C / N0) of each frequency band is calculated in real time, retaining only data from bands with a signal-to-noise ratio above a set threshold. This filters out noise interference and improves data quality. This reduces invalid data transmission, lowers the cloud processing burden, and shortens response latency.

[0126] Dual-channel transmission link (4G network and Beidou short message): This mode uses a backup transmission mode of 4G high-speed transmission and Beidou short message satellite communication. When the 4G signal is weak or network latency is high, it automatically switches to Beidou short message transmission for compressed data packets to avoid communication interruption. The cloud reassembles the dual-channel data through timestamp alignment and quality identifier verification to ensure data transmission continuity.

[0127] Local Area Augmentation Network and Multi-Base Station Joint Error Modeling: Utilizing a cluster of reference stations to build a regional differential network, a system is used to correct satellite signal propagation errors in real time. Ionospheric delay, tropospheric refraction, and multipath effects are jointly modeled using multi-base station data, dynamically generating error compensation parameters. Baseline calculation and network adjustment algorithms eliminate positioning residuals, enabling millimeter-level 3D coordinate resolution.

[0128] Spatiotemporal reference data stream: A unified data sequence that has undergone error correction and time synchronization. High-precision timestamps are used to align multi-source data (such as satellite displacement and strain measurements), and elevation component deviations are secondary corrected using atmospheric refraction compensation parameters. This ensures spatiotemporal consistency for subsequent data fusion and prevents cumulative errors.

[0129] Kalman filter cross-dimensional data fusion: This state-space model-based filtering algorithm fuses satellite displacement and strain sensor data. This eliminates noise from individual sensors and extracts key deformation parameters such as horizontal displacement, vertical settlement, and crack propagation rate. Incorporating temperature and seepage field data enhances the model's adaptability to multi-physics coupling effects.

[0130] Global Deformation Dynamic Model: This hybrid model integrates finite element analysis and data-driven analysis to simulate the deformation response of dam structures under multiple loads. It integrates multiple parameters such as displacement, strain, temperature, and seepage pressure. Deformation monitoring indicators such as displacement, velocity, and acceleration are used to assess structural stiffness degradation trends.

[0131] 3D deformation heatmaps and digital twin mapping: Digital twin technology visualizes deformation data and identifies high-risk areas. Exceeding displacement limits are coded using color gradients and overlaid with transient response consistency verification results to generate a localizable risk report. This assists inspection terminals in quickly locating hazardous areas, improving response efficiency.

[0132] Closed-loop adaptive mechanism: The system dynamically optimizes key parameters based on historical data and feedback, including monitoring point density, signal-to-noise ratio screening thresholds, and atmospheric refractometer compensation parameters. Machine learning algorithms (such as LSTM) are used to predict environmental parameter trends, ensuring long-term stability in monitoring accuracy.

[0133] The specific implementation of the present invention is as follows: An array of Beidou monitoring stations and reference stations are deployed in the dam and slope areas of a pumped-storage power station, using a multi-mode GNSS receiver to synchronously capture multi-band satellite signals from the Beidou, GPS, GLONASS, and Galileo systems. The monitoring stations integrate edge computing units that use signal-to-noise ratio threshold screening technology to remove interference from low-signal-to-noise ratio frequency bands in real time, retaining valid frequency band data for weighted fusion to generate pre-processed satellite observation data with signal quality identifiers. The pre-processed data is uploaded to a cloud server in real time via a preset 4G network and Beidou short message dual-channel transmission link. The Beidou short message transmission mode is automatically enabled when the signal strength falls below a preset threshold, resolving the problem of data transmission interruption in communication blind spots caused by complex terrain and ensuring data real-time and integrity.

[0134] A cloud-based server constructs the Beidou local area augmentation network. Based on the spatial distribution characteristics of the base station cluster, a multi-base station joint error modeling approach is used to correct for ionospheric delay and tropospheric refraction errors in real time. The base stations' built-in rubidium atomic clocks and satellite common-view modules achieve sub-nanosecond time synchronization. This combines baseline solution with an anti-error network adjustment algorithm to eliminate positioning bias and generate a time-space reference data stream with millimeter-level accuracy. This data stream undergoes secondary calibration using an atmospheric refraction compensation module. This module uses a preset meteorological parameter library to dynamically match signal propagation correction coefficients to real-time environmental conditions, eliminating systematic measurement errors caused by atmospheric environmental changes and improving the measurement reliability of elevation components.

[0135] The calibrated spatiotemporal benchmark data stream is timestamped with strain data collected by a static level and multi-point extensometers. A Kalman filter algorithm is used to cross-dimensionally fuse displacement and strain features to extract horizontal displacement, vertical settlement, and crack propagation rate parameters. This fused data, combined with temperature and seepage sensor data, is fed into a global deformation dynamic model. A hybrid finite element and data-driven architecture outputs displacement, velocity, and acceleration monitoring parameters. The early warning analysis module dynamically sets multi-dimensional thresholds based on a stiffness degradation model. When monitored values ​​continuously exceed limits, a graded early warning signal is triggered. The digital twin model maps the exceeded areas to a three-dimensional thermal map, overlaying transient response consistency verification results to mark high-risk coordinates. The monitoring point density and data collection frequency are dynamically optimized based on historical feature distribution. Warning thresholds are updated based on feedback from inspection terminals, achieving closed-loop adaptive monitoring. A dual-channel transmission link switches transmission modes based on network status. A cloud server uses redundancy checks and linear interpolation to supplement low-quality data segments, ensuring data integrity and timeliness. Ultimately, this module enables millimeter-level dynamic monitoring of the dam's entire deformation and provides real-time safety warnings.

[0136] The present invention solves the problem of insufficient manual observation frequency through multi-system satellite signal fusion and dynamic error correction technology. Beidou monitoring stations and reference station arrays are deployed in the dam body and slope areas to synchronously receive multi-band satellite signals and perform edge computing processing, and use the signal-to-noise ratio threshold to filter and fuse effective frequency band data. Pre-processed data is uploaded to the cloud server in real time through a preset dual-channel transmission link, where the Beidou short message mode is automatically enabled when the signal strength is insufficient to ensure data transmission continuity under complex terrain. The local area enhancement network is constructed based on a group of reference stations, and a multi-base station joint error modeling method is used to correct ionospheric and tropospheric errors in real time. It is combined with a high-precision time synchronization device to eliminate positioning deviations, generate a millimeter-level precision time-space reference data stream, and realize unified time-space reference processing of multi-source data.

[0137] The problem of inconsistent spatiotemporal benchmarks is resolved through multi-physics field coupling modeling and real-time deformation analysis technology. The calibrated spatiotemporal benchmark data stream is spatiotemporally aligned with the strain data from the static level and multi-point displacement gauge. The Kalman filter algorithm is used to cross-dimensionally fuse the displacement and strain data to extract horizontal displacement, vertical settlement, and crack propagation characteristics. The fused data is combined with temperature field and seepage field parameters to input the global deformation dynamic model. The transient response consistency is verified through time series analysis, and the displacement, rate, and acceleration monitoring parameters are output. The digital twin model maps the coordinates of the out-of-limit area to a three-dimensional deformation thermal map, superimposes the consistency verification results to mark high-risk areas, and realizes the dynamic visualization and precise positioning of deformation characteristics.

[0138] A closed-loop adaptive mechanism solves the problem of not being able to obtain millimeter-level three-dimensional deformation data in real time. The warning threshold is dynamically set according to the stiffness degradation model. When the monitoring value exceeds the limit continuously, a graded warning signal is triggered, and the threshold parameters are optimized through feedback from the inspection terminal. The distribution density of monitoring points and the frequency of data acquisition are dynamically adjusted according to historical monitoring characteristics, and the signal processing parameters are optimized in real time based on data quality assessment indicators. Atmospheric refraction compensation parameters are regularly updated in combination with environmental change data to ensure the consistency of measurement accuracy under complex meteorological conditions. The dual-channel transmission link switches adaptively according to the network status, and redundant check and interpolation completion technology eliminates the impact of low-quality data, forming a closed-loop management of the entire process from data acquisition to risk disposal, thereby improving the real-time and reliability of dam safety monitoring.

Claims

1. A pumped storage power station dam safety monitoring method based on Beidou positioning is characterized by: include: Step 1: Receive multi-band satellite signals, including Beidou satellite signals, GPS satellite signals, GLONASS satellite signals, and Galileo system satellite signals, perform edge computing on the received multi-band satellite signals, filter and fuse effective frequency band data through a signal-to-noise ratio threshold, generate pre-processed satellite observation data, and establish a dual-channel transmission link through a preset 4G network and a preset Beidou short message to upload the data to a preset cloud server in real time; Step 2: Build a local augmentation network on the cloud server, perform dynamic error correction on the uploaded pre-processed satellite observation data, generate three-dimensional coordinate solution results, and output the calibrated spatiotemporal reference data stream; Step 3: performing spatiotemporal alignment on the satellite displacement data in the calibrated spatiotemporal reference data stream with the strain measurement data collected by the static level and the strain measurement data collected by the multi-point displacement meter. By fusion of the data, the transient response characteristics of the displacement and strain are extracted, and a global deformation dynamic model that integrates the coupled effects of the temperature field and the seepage field is constructed. Step 4: Input the satellite displacement data of the spatiotemporal reference data stream and the strain measurement data into the global deformation dynamic model, and output the displacement, velocity, and acceleration monitoring parameters of the dam structure; Step 5: Setting warning thresholds corresponding to the displacement, velocity, and acceleration monitoring parameters of the dam structure on the cloud server. When the displacement, velocity, and acceleration monitoring parameters of the dam structure exceed the corresponding warning thresholds, generating a graded warning signal including a risk level and location coordinates based on the values ​​of the exceeded parameters; Step 6: constructing a three-dimensional deformation thermal map based on the displacement, velocity, and acceleration monitoring parameters of the dam structure, and pushing a risk report including the thermal map coordinate location to a pre-configured terminal device via a pre-set communication link; The step 1 comprises: Beidou monitoring stations and reference stations deployed in the dam and slope areas synchronously receive the Beidou, GPS, GLONASS and Galileo satellite signals; Perform edge computing processing on the received multi-system satellite signals, filter through signal-to-noise ratio thresholds to retain valid frequency band data, and add signal quality identifiers; Predicting a communication blind spot based on a signal quality identifier, and switching to a Beidou short message transmission mode in the dual-channel transmission link when the signal strength is detected to be lower than a preset strength threshold; Upload the pre-processed satellite observation data to the cloud server through the preset 4G network and Beidou short message dual channel; Perform spatiotemporal alignment on the pre-processed data in the cloud server to generate a continuous data stream that is input into step 2 to perform dynamic error correction. The step 2 includes: A BeiDou local augmentation network is constructed by deploying base stations, and a multi-base station joint error modeling method is used to correct ionospheric and tropospheric errors in real time for pre-processed satellite observation data. Sub-nanosecond time synchronization is achieved based on the high-precision time synchronization device preset at the base station, and positioning deviation is eliminated through the combined processing of baseline solution and network adjustment; Input the corrected satellite observation data into the three-dimensional coordinate solution processing flow to generate a spatiotemporal reference data stream with a timestamp; Performing secondary calibration on the space-time reference data stream using preset atmospheric refractometer compensation parameters, and outputting the calibrated space-time reference data stream; The step 3 comprises: extracting satellite displacement data from the calibrated spatiotemporal reference data stream, and performing time stamp alignment with the strain data collected by the static level and the strain data collected by the multi-point displacement meter; The preset Kalman filter algorithm is used to perform cross-dimensional fusion of the aligned displacement and strain data to extract horizontal displacement, vertical settlement and crack propagation characteristics; The fused data is combined with the field data collected by the temperature sensor and the seepage pressure gauge and input into the global deformation dynamic model; Verify the transient response consistency between the fused data and the global deformation dynamic model output through time series analysis, and generate transient response consistency verification results for early warning rule calls; The step 4 includes: establishing a stiffness degradation model on a cloud server based on the displacement parameters output by the global deformation dynamic model and setting multi-dimensional warning thresholds for deformation, velocity, and acceleration; Compare the displacement monitoring value of the deformation dynamic model with the warning threshold in real time, and trigger a graded warning signal when the monitoring value exceeds the limit for a set number of consecutive times; The coordinates of the out-of-limit area are mapped to the 3D deformation thermal map through the preset digital twin model, and the transient response consistency verification results are superimposed to obtain the marked high-risk areas. The marked high-risk areas are pushed to the preset inspection terminal, and the warning threshold parameters are updated based on the disposal results fed back by the inspection terminal.

2. The method for monitoring dam safety of a pumped storage power station based on Beidou positioning according to claim 1 is characterized in that: The step 5 comprises: Collect real-time displacement monitoring information based on the deployed base stations, and eliminate system errors caused by base station structural deformation through a preset dynamic coordinate update algorithm; Based on the historical monitoring feature distribution law generated by the spatiotemporal alignment and data fusion processing, the spatial distribution density of the initially deployed monitoring points and the data acquisition frequency are dynamically optimized, wherein the historical monitoring feature distribution law is obtained by the time series analysis of the transient response consistency verification results; Dynamically optimize the noise suppression coefficient and multi-band signal weight matching parameters in the signal-to-noise ratio threshold screening process based on the data quality evaluation index generated by the secondary calibration process of the atmospheric refractometer compensation parameter; Combined with the environmental parameter change data in the graded warning signal, the atmospheric refractometer compensation parameters described in step 3 are updated according to the preset calibration cycle.

3. The method for monitoring the safety of a pumped storage power station dam based on Beidou positioning according to claim 2 is characterized in that: The step 6 comprises: When the 4G network signal strength is higher than the set signal strength threshold, the pre-processed satellite observation data generated in step 1 is uploaded through the high-speed channel; When it is detected that the network delay exceeds the preset tolerance threshold or the data packet loss rate increases, switching to the Beidou short message transmission mode in the dual-channel transmission link to send compressed data packets; The cloud server performs redundancy check and timestamp alignment on the dual-channel transmission link data based on the signal quality identifier added in step 2; The reorganized data is transmitted to the spatiotemporal reference data stream update processing flow described in step 3, and a linear interpolation completion operation is performed on the low-quality data segments marked by the identifier.

Images